Genetic Optimization of Hybrid Load Balancing Algorithm for Grid Computing Environment

Recent advances in internet technology has lead to the formation of distributed computing environment, geographically spreading over the globe for solving high-end computational tasks. To increase the potential of such large-scale distributed systems, it requires proper load-balancing algorithm arises. Since the problem of Load Balancing is NP-Complete, therefore it requires heuristic or approximation algorithm. In this paper, we have conducted an experimental verification of Genetic Optimization of the Hybrid Load Balancing Algorithm, by considering Value Function as optimization parameter. With throughput as performance parameter, we have evaluated the algorithm on a grid based on MetaCentrum Project (Czech Republic National Grid Infrastructure). Further, we have used a Check Point Based Recovery Technique in establishing the reliability of the grid.